Confinement Effects On High-Strength Concrete: ACI Structural Journal July 2001
Confinement Effects On High-Strength Concrete: ACI Structural Journal July 2001
net/publication/292649814
CITATIONS READS
53 1,254
4 authors, including:
Hamid Saadatmanesh
The University of Arizona
77 PUBLICATIONS 6,206 CITATIONS
SEE PROFILE
All content following this page was uploaded by Panos D. Kiousis on 12 March 2016.
An experimental investigation was conducted on the stress-strain 4. The split cylinder strength for low-strength concrete
characteristics of steel sleeve confined high-strength concrete (LSC) and HSC can be conservatively represented by the ex-
(HSC). The axial load and strains of concrete, and the axial and pression fcp′ = 6 f c ′ in psi units;
hoop strains of the confining steel sleeves were measured. From 5. The ACI Code1,2 current expression for the modulus of
these measurements, accurate stress-strain relations of the con- rupture fr = 7.5 f c ′ may be too conservative for HSC. The
crete core were produced, along with confinement calculations
alternative expression fr = 2( f c ′ )2/3 appears to be more rep-
based on von-Mises elastoplastic response of the steel sleeves.
Confinements ranging from 5 to 19 MPa were calculated. This con- resentative of the test data;
finement had a profound effect on the strength of concrete, as much 6. In the inelastic range, HSC exhibits less volume dilata-
as tripling its unconfined strength of 70 MPa. The increase in duc- tion, which explains the observed reduced confining effec-
tility was found to develop slower for low amounts of confining tiveness of spiral reinforcement;
steel due to a lagging development of confining pressure. This was 7. The effects of high strain rate on the strength appear to
attributed to the reduced tendencies of HSC to exhibit lateral be less in HSC;
expansion and the early yielding of the sleeve hoop stresses due to 8. At the material level, HSC is less ductile than NSC. At
the biaxial nature of stresses within the sleeves. the section level for reinforced concrete elements, however,
the effects of the reduced ductility are less clear; and
Keywords: confinement; ductility; high-strength concrete. 9. It may be possible to obtain an essentially flat descend-
ing region of the stress-strain curve, provided that spirals
INTRODUCTION with sufficiently high yield strength and low pitch are used.
Concrete structures in areas of significant earthquake Research reported by Yong, Nour, and Nawy5 Mander,
overloads require increased strength and ductility, and accu- Priestley, and Park6 and other researchers shows trends sim-
rate methods to estimate them. Improvements in the strength ilar to the trends previously reported.
and ductility of concrete structures are commonly obtained The reduced effectiveness of the transverse reinforcement
by increased lateral confinement with the use of transverse caused by the lack of extensive microcracking, and the over-
reinforcement. Spirals and ties in the shape of circles, all brittleness of HSC, indicated by the steep prefailure as-
squares, diamonds, octagons, overlapping squares, and cent followed by the very rapid postfailure descent of the
others are being used for this purpose. In addition, numerous stress-strain relation, are probably the most important fea-
computational approaches have been presented in the litera- tures that affect the way one should design using HSC.
ture to estimate the effects of confinement on the strength Martinez, Nilson, and Slate7 found that the common ap-
and ductility of concrete. Some basic concepts related to proach of estimating lateral confinement, by assuming
strength and ductility have been introduced in the ACI that the transverse reinforcement has yielded, results in un-
Code,1,2 especially with respect to the ductility of compres- conservative estimation of the load capacity and ductility,
sion members. These concepts are empirical or semi-empir- especially when high-strength steel (> Grade 60) is used.
ical in nature, and have been based on large numbers of tests Nevertheless, recent research in Japan8,9 indicates that HSC
on low or medium strength concrete, while high-strength can be confined effectively even though such confinement is
concrete (HSC) has been mostly overlooked. probably obtained without necessarily achieving yielding of
the transverse reinforcement.
The applications of HSC are constantly increasing due to The intent of this study is to investigate the development
the rising demands for stronger and more inexpensive struc- of confinement in laterally reinforced HSC, and to develop a
tural materials. Recent research has shown that HSC behav- simple model to predict the effect of such confinement.
ior deviates from that of normal-strength concrete (NSC). Transverse reinforcement in the form of steel sleeves of dif-
Thus, classical approaches to design may not be as effective ferent wall thicknesses has been chosen to facilitate the de-
or as safe when they are applied to HSC. velopment of uniform lateral stresses. This process can also
Related studies by Ahmad and Shah3,4 indicate that: serve as a small scale testing of HSC columns that are reha-
1. An increase in compressive strength decreases the con- bilitated by steel wrapping to improve earthquake resistance.
fining effectiveness of spiral reinforcement;
2. There is a significant difference in the compressive RESEARCH SIGNIFICANCE
stress-strain curves of NSC and HSC. The stress-strain be- Literature on experimental studies on HSC, especially in
havior of HSC is linear to a higher fraction and postpeak the material level, is relatively scarce. The existing design
stress loss is more abrupt;
3. The ACI equation for estimating the secant modulus of ACI Structural Journal, V. 98, No. 4, July-August 2001.
elasticity Ec = 33w1.5 f c ′ predicts values as much as 20% MS No. 00-185 received August 9, 2000, and reviewed under Institute publication
policies. Copyright © 2001, American Concrete Institute. All rights reserved, includ-
too high for concrete with a compressive strength in the ing the making of copies unless permission is obtained from the copyright propri-
etors. Pertinent discussion will be published in the May-June 2002 ACI Structural
vicinity of 12,000 psi (~80 MPa); Journal if received by January 1, 2002.
the height of the sleeve was increased to 216 mm (8.5 in.), as Table 2—Mixture proportions for 1 yd3 of high-
shown in the schematic of Fig. 1. strength concrete (HSC)
Component Weight, lb
Material properties Cement Type I and II, portland 801
The specimens were prepared in the Structural Engineer- Force 10,000 microsilicate 64
ing Laboratory of the University of Arizona and the material
Water 288
properties were obtained as follows:
Sand 1269
Steel sleeve—The sleeve is DOM 1020/ASTM A 513 with
yield stress σy = 448 MPa (65,000 psi) and ultimate strength Crushed rock 1615
σu = 552 MPa (80,000 psi). The elongation of steel at failure Admixture WRDA 19 (high-range water-reducing
admixture) (ASTM C 94 Type A and F)
8
is 10%. The modulus of elasticity is E = 195 GPa (28,250 ksi).
WRDA 64 (H2O reducer) 1.5
These values were obtained at the Rock Testing Laboratory of
the University of Arizona by building a cylindrical chamber
using a steel sleeve and two spherical ends. The chamber was while the axial stresses σa were integrated over the core sur-
loaded by internal hydraulic pressure to failure. face and subtracted from the applied load.
Dry lube—The spray lube used in the inner surface of the All tests were performed using the 2.2 MN (500,000 lb)
steel sleeve to reduce shear transfer was of type 2W757B universal compression machine at the Rock Testing Labora-
T.F.E. tory at the University of Arizona. The specimens were
placed in the loading frame (Fig. 3) and were loaded at a
Concrete—The mixture proportions for one cubic yard of
strain rate of 0.00031/s.
HSC are outlined in Table 2. The expected compressive
strength fc′ was 76 MPa (11,000 psi). The measured com-
pressive strength was 71 MPa (10,300 psi) at 28 days. All Test results
specimens, confined and unconfined, were tested within a The axial stress-strain relations of the concrete core are
period of two days. The measured unconfined strength of presented in Fig. 4 as a function of the volumetric percent of
71 MPa is assumed for all specimens. The fineness modulus confining steel, which is calculated as 4t(D + t)/D2. The ef-
of coarse aggregate and sand were 5.71 and 3.39, respectively. fects of confinement are very apparent both in increasing the
axial strength, and in increasing the ductility of the speci-
Concrete mixing—The mixing process to obtain the HSS
mens. Also noticeable is the increase of the strain at peak
is as follows:
stress from approximately 0.002 for the unconfined speci-
1. Mix rock and microsilicate for 3 to 5 min; mens to almost 0.01 for the heavily confined specimen (steel
2. Add 75 to 80% water and WRDA 64 and mix for 3 min; wall thickness = 3.1 mm or 12.3% confining steel ratio). The
3. Add sand and cement. Mix for 2 min; set 3 min; mix an specimens with sleeve wall thicknesses up to 1.16 mm (or
additional 2 min; 4.6% confining steel ratio) were tested to failure, which oc-
4. Add WRDA 19 and mix for 3 min; and curred when the sleeve raptured. This is clearly indicated in
5. Let the mixture set for 3 min and place in the molds. Fig. 5 for specimens with confining steel of 1.5, 3.1, and
4.6% (1B, 2B, and 3B, respectively). At that time, shear fail-
Instrumentation and test setup ure plains inclined to 65 to 70 degrees were formed. Loading
Strain gages Type WA-06-500WT-120 were used to mea- of the other specimens with the larger confinement was ter-
sure the hoop and axial strains of the outer surface of the minated before ultimate failure due to deformation limita-
steel sleeves at two diametrically opposite points at the mid- tions of the testing machine. In some cases, however, shear
height of each specimen. The vertical strains on the HSC failure plains seemed to begin to form.
core were measured by direct deformation measurements To evaluate the actual stresses that the steel sleeve applies
using linear variable displacement transducers (LVDTs). on the concrete core, the axial εas and hoop εθs strains of the
The incremental steel strains were used to evaluate the steel sleeve were recorded and are plotted in Fig. 6. As pre-
corresponding stresses. The initial yield of steel was calcu- viously stated, a dry lube was used in the inner surface of the
lated using the von-Mises criterion of failure followed by steel sleeve to reduce the shear transfer between the concrete
complete integration of equations of plasticity to evaluate core and the sleeve. As can be seen from Fig. 6, the axial
the postyield response. The evaluated hoop stresses σθ were strains of the steel sleeve were not eliminated. Comparing
used with Eq. (1) to evaluate the applied confinement p, Fig. 4 and 6, it is concluded that, at best, a 50% reduction was
ksi), yield stress σy = 448 MPa (65 ksi), and Poisson’s ratio
ν = 0.3. The plasticity equations and the numerical imple-
mentation of the von-Mises criterion are well known and are
not discussed herein. The results of this analysis, however,
are very significant for the purpose of this study and are
shown in Fig. 7 as a plot of the vertical strain of the concrete
core versus the steel sleeve induced confinement. The signif-
icance of these results is quite evident. It is found herein that:
1) the confining pressure is smaller than that implied from
Fig. 4—Effects of steel confinement in strength and ductility
the assumption that σθ = σy for the steel sleeve (compare the-
of concrete cylinder.
oretical and experimental confinement values in Columns 3
and 4 of Table 1); and 2) that confinement develops in a lag-
achieved. In this figure, as is done throughout this paper, ten- ging way.
sile strains are negative.
The reason for the lower value of the confining pressure is
illustrated in Fig. 8, which shows the von-Mises yield func-
Analysis of test results tion under plane stress conditions. A typical wall element of
As was previously stated, the selection of the sleeve wall the steel sleeve is subjected to such biaxial state of stress
thicknesses was based on the assumption that the hoop stress with a compressive axial stress σa and a tensile hoop stress
of the sleeve σθ = σyield. It was expected that this assumption
σθ. This results in a stress path in the fourth quadrant of the
would not be accurate, but was used as a first estimation. To
σa − σθ space, which at failure results in both σa and σθ to
calculate the actual confinement stresses on the concrete
be less than σy. This of course implies that the steel sleeve
core, the strain history of the steel sleeve was integrated
using the von-Mises criterion for steel. No hardening was carries part of the axial load or
assumed for the steel. The material parameters for steel are
as follows: modulus of elasticity E = 200,000 MPa (29,000 P = Pc + P s (3)